Nitin Agrawal is the chief financial officer of CoreWeave, the executive who owns the number every analyst is arguing about this week: how much of the company's $99.4 billion remaining performance obligations backlog is real revenue on the way and how much is a promise still waiting on GPUs to arrive. CoreWeave reports second-quarter 2026 results Tuesday, August 11, after the close, with the call at 5:00pm ET, and Agrawal is the person who will explain the guidance behind those numbers. Nothing about Tuesday has happened yet, and this article treats it that way throughout.
CRWV traded around $91.78 in Monday's pre-market quote, a level rather than a session move, since every stock percentage on a live feed right now compounds Friday's close across the weekend. A separate piece on this site runs the backlog-conversion math itself for Tuesday's print. This one answers a different question: who is delivering that guidance, and what does his record say about how to weigh it. His resume covers finance at Amazon, Google and now CoreWeave, three different versions of the same cloud-computing business, and that pattern is the most useful lens for reading Tuesday's call.
Why Agrawal Is the Person to Watch Before Tuesday's Numbers Land
A CFO does not set demand for GPU capacity, but a CFO decides how a company talks about it, and that is exactly what makes Tuesday's call about Agrawal as much as it is about CoreWeave. Every figure investors are waiting on, from an updated backlog total to the pace of margin expansion, gets framed and contextualized by him on the call before analysts ever get to ask a question. CoreWeave's CEO Michael Intrator, whom this site has profiled separately, sets the strategy and signs the customer deals. Agrawal is the one who has to explain, in real time, if that strategy is converting into cash on the schedule the stock price assumes.
That split matters this quarter. CoreWeave's Nasdaq-100 inclusion earlier this year pulled in a wave of index-fund ownership that had never had to evaluate a neocloud's finance chief before, and Tuesday is the first quarterly test of how much that broader investor base trusts the numbers Agrawal presents. The stakes are not abstract. A company sitting on a backlog many multiples larger than its current quarterly revenue lives or dies, in the market's eyes, on how well its CFO can explain the conversion timeline under scrutiny.
The Career Path That Built a Cloud-Finance Specialist
Agrawal has been CoreWeave's CFO since March 2024, confirmed by both the company's own investor relations announcement and its current leadership page. Before that, his resume is a tour through the finance function of cloud computing at three different scales.
|
Company
|
Role
|
Dates
|
|
CoreWeave
|
Chief Financial Officer
|
March 2024 to present
|
|
Google Cloud
|
Vice President, Finance
|
May 2021 to March 2024
|
|
Mapbox
|
Chief Financial Officer
|
August 2019 to April 2021
|
|
Amazon Web Services (Compute Services)
|
Finance Director
|
January 2015 to July 2019
|
|
Microsoft
|
Finance group manager
|
roughly seven years, reported
|
He holds a B.Tech in Engineering from the National Institute of Technology in Kurukshetra, India, and an MBA in Finance from Duke University's Fuqua School of Business, both confirmed on CoreWeave's own leadership and investor relations pages.
Two of those rows carry a note worth flagging rather than smoothing over. CoreWeave's current official bio lists the AWS role as Finance Director of the Compute Services division, the title used above, though the company's own 2024 announcement press release called the same job "CFO of the Compute Services division." Both are CoreWeave's own language, just from different years. The Microsoft row is the thinnest of the six, absent from CoreWeave's current leadership and investor bio pages and traceable instead to trade press coverage of the 2024 appointment plus a secondary aggregator placing that tenure at roughly 2008 to 2015, a span that also works out to about seven years. Every other row is confirmed directly by CoreWeave's own leadership page and investor relations bio, as close to a primary source as a corporate biography gets.
What a Cloud CFO Actually Manages
A cloud CFO's job looks nothing like a retailer's or a bank's, and the difference is the fastest way to understand what Agrawal is managing heading into Tuesday.
The first job is committed capacity. A cloud provider does not sell a product off a shelf. It sells access to compute that has to be physically built, powered and connected before a single dollar of the contract can be recognized as revenue. When CoreWeave signs a multi-year GPU deal, that commitment shows up on the books as a promise long before the data center behind it is finished, which is exactly why the backlog number is so much larger than current quarterly revenue.
The second job is GPU depreciation scheduling, and this one moves reported profit more than almost anything else on the income statement. A depreciation schedule is the assumption a company makes about how many years a chip will keep earning money before it is written down to zero on the books. Stretch that assumption from three years to six, and the same GPU purchase cuts the annual depreciation expense roughly in half, which lifts reported operating income without changing a single thing about the underlying business. The economics of AI chip supply and useful life are already a live debate across the sector, and every neocloud CFO, Agrawal included, is making a judgment call on this assumption that shapes how profitable the company looks quarter to quarter.
The third job is why remaining performance obligations is the metric that actually matters. Remaining performance obligations, or RPO, is the total dollar value of contracted work a company has signed but not yet delivered and billed. For a business that sells capacity years in advance, RPO functions as the closest thing to a demand signal that exists, because current-quarter revenue only reflects what has already been built and switched on, not what customers have already agreed to buy. That is why Q1's $99.4 billion RPO figure gets more attention on Tuesday's call than the quarter's actual revenue print. It is the number that tells investors if demand is still outrunning CoreWeave's ability to build, or if the growth story is starting to catch up with itself.
Three Generations of the Same Cloud Playbook
The detail that makes Agrawal worth a profile rather than a one-line executive bio is that he has sat inside cloud computing's finance function at three different points in its evolution. At Amazon Web Services, he managed finance for Compute Services during the years the hyperscale utility model matured into a predictable, high-margin business. At Google Cloud, he moved to a division still proving it could turn a profit, and multiple reports credit him and his team with helping drive it toward profitability while building its AI and machine-learning investment strategy. At CoreWeave, he is now the finance chief of a specialist that skipped the general-purpose cloud model and built around GPU capacity from day one, a company still spending far more on capex than its backlog math would suggest it needs on a straight-line basis.
That progression is not a coincidence, and it is the closest thing to a real edge a CFO can bring to this role. Someone who has only ever worked inside a mature, profitable cloud business tends to underestimate how much capacity-building costs distort the numbers early on. Someone who has only worked inside a capacity-constrained startup tends to lack a reference point for what the business looks like once growth slows and margins have to hold up on their own. Agrawal has watched both stages up close, at two of the largest cloud operators in the world, before landing at a company still living through the first one.
The pattern also has a useful comparison outside CoreWeave. Oracle's own cloud infrastructure business has drawn similar scrutiny over the past year for a backlog that dwarfs its current quarterly cloud revenue, built on the same logic of signing capacity commitments ahead of the data centers that will fulfill them. It is not the same company or the same balance sheet, but it is the same underlying question investors are asking of both finance teams. Does a backlog that size represent demand the company can actually build into, or a number that looks better on a slide than it will look on a cash flow statement three years from now.
How to Read His Guidance on Tuesday
None of Tuesday's outcomes are known yet, so the useful exercise is building the framework for reading what Agrawal actually says rather than guessing at it in advance.
Listen for specificity over repetition. CoreWeave's Q1 guidance said roughly 36% of backlog converts within two years and about 75% within four. If Agrawal repeats those exact figures without any update, that is a company holding its prior line rather than sharpening it. If he narrows the range, adds a near-term conversion figure, or ties a portion of the backlog to a named customer ramp, that additional precision is itself informative, regardless of which direction the number moves.
Listen for how he frames new backlog additions against capacity delivery. Cantor Fitzgerald's Brett Knoblauch has modeled roughly $40 billion in new Q2 backlog additions, matching the pace CoreWeave set in Q1. How Agrawal discusses new signings, tied to a specific data center or power delivery timeline rather than left as a standalone dollar figure, signals if the finance team is anchoring growth claims to physical build-out or leaving that connection for analysts to assume.
Listen for the capex language against the $31-35 billion 2026 guidance. Wells Fargo's Michael Turrin models operating margin climbing from roughly 3% in Q2 toward 13% by Q4 2026, an estimate that is his own and not CoreWeave's guidance. How closely Agrawal's commentary on capex pacing and margin trajectory tracks that model, or departs from it, is the kind of detail a CFO chooses his words around carefully, and it is worth listening for exactly that choice.
Frequently Asked Questions
Who is the CFO of CoreWeave?
Nitin Agrawal has been CoreWeave's chief financial officer since March 2024. He previously served as vice president of finance for Google Cloud, CFO of Mapbox, and finance director for Amazon Web Services' Compute Services division, giving him finance leadership experience across three different cloud computing companies before joining CoreWeave.
When did Nitin Agrawal become CoreWeave's CFO?
Agrawal joined CoreWeave as chief financial officer in March 2024, confirmed by both the company's investor relations announcement at the time and its current leadership page. He arrived directly from Google Cloud, where he had spent almost three years as vice president of finance.
What companies did Nitin Agrawal work for before CoreWeave?
Before CoreWeave, Agrawal was vice president of finance at Google Cloud from 2021 to 2024, chief financial officer of the mapping company Mapbox from 2019 to 2021, and finance director of Amazon Web Services' Compute Services division from 2015 to 2019. Earlier reporting also places roughly seven years of finance roles at Microsoft before his AWS tenure began.
What is CoreWeave's remaining performance obligations backlog?
CoreWeave reported $99.4 billion in remaining performance obligations as of its first-quarter 2026 results, published around May 7, 2026. That figure represents contracted revenue the company has signed but not yet recognized because the underlying compute capacity has not been fully built and delivered, and no updated Q2 figure exists until Tuesday's report.
Bottom Line
Nitin Agrawal's record across Amazon, Google and CoreWeave is the reason his framing of Tuesday's numbers deserves more attention than a typical earnings-call CFO segment usually gets. If his commentary sharpens the 36%-in-two-years backlog conversion figure with new specifics rather than repeating it verbatim, treat that as a finance team getting more confident in its own math, not less. If new backlog additions land near Cantor's $40 billion estimate and Agrawal ties them to a named capacity timeline, that supports the reading that demand is still outrunning what CoreWeave can build. If backlog growth falls well short of that estimate, or if the conversion ratios get walked back without a clear explanation, that shifts the story from a supply constraint to a demand question, which is a materially different setup for the stock. Nobody outside CoreWeave knows which of those happens Tuesday at 5:00pm ET, but his career says he has seen this exact tension play out twice before, at two of the largest cloud businesses on earth.
This article is for informational purposes only and does not constitute financial or investment advice. Cryptocurrency and equity trading involves substantial risk. Always conduct your own research before making trading decisions.
